Is Microsoft a Dying Business? The Real Reason It's Down 20% This Year

Let's Answer the Market's Question First

I don't think Microsoft is a dying business. The stock is down more than 20% this year, yet cloud revenue is still growing 30% and operating margins have actually risen over the past decade. The problem isn't the company — it's that the market hasn't figured out how to price this transition yet.

Microsoft has badly lagged the market this year. It's down more than 20% year to date while the market as a whole is up 10-13%. That's a gap of more than 30 percentage points. When a large, high-quality name diverges that much, I reflexively ask two questions: is the fear justified, or has the market gotten ahead of itself?

The Two Things the Market Fears

Two fears are pressing on the stock.

The first is money. Microsoft is pouring a staggering $190 billion into AI infrastructure. The number is so large that investors can't help asking, “So when does this actually pay off?” Right now, none of that payoff is reflected in the price.

The second — and this is the bigger one — is the worry that AI agents will cannibalize the traditional software business. Microsoft makes enormous money from subscription software like Microsoft 365 and Teams. Businesses pay for every seat, every license. But if AI can do those tasks automatically and more cheaply, why would a company keep buying all those seats? That fear is hammering the entire software industry right now, Microsoft included.

The Numbers Tell a Different Story

Open the actual results, though, and this is not the picture of a dying company.

In Q3 of fiscal 2026 (which ended in March), total Microsoft Cloud revenue reached $54.5 billion, up 30% year over year. On top of that, the company's dedicated AI business surpassed an annual revenue run rate of $37 billion. On the consumer side, Office 365 grew 7% year over year; on the commercial side, more than 450 million users with 17% growth.

The quality of profits is even more striking. A ten-year average operating margin of 33.9%, five-year 36.7%, and 39.34% over the last year. Across a decade the margin climbed from the low 33s to nearly 40. Return on capital has held around 21% over five years. Looking at this table, I kept asking myself whether these are the numbers of a declining business. Nobody knows the future — but these numbers don't belong to a dying company.

So, What's It Worth?

Start with Microsoft's “price.” Market cap is $2.8 trillion; enterprise value is $3.02 trillion. That $200 billion difference is net debt — and last year the company generated $73 billion in free cash flow, even after heavy capital spending. It can comfortably carry that debt.

On valuation, it trades at 38 times free cash flow and 22 times earnings. On the eight pillars I use, only two flags come up for Microsoft — both valuation-related. In other words, “it's expensive.” But remember this: the same 22x earnings and 38x cash flow would be absurdly cheap if the company grows 100% a year, and absurdly expensive if it shrinks 3% a year. A multiple only means something paired with a growth rate.

Analyst estimates have earnings per share going from $17 to $40 over seven years — about 2.5x. Revenue more than doubles, from $336 billion to $760 billion. That's over 10% growth a year.

I ran my own assumptions on a 10-year basis: revenue growth of 7/10/13%, operating margins of 34/37/40%, and a 10-year exit P/E of 20/23/26 — a premium I'm willing to give a high-quality business with a wide moat. Requiring just a 9% return with no margin of safety, intrinsic value comes out at $360 on the low end, $820 on the high end, and $545 in the middle — roughly a 14% expected return at today's price. Plug in my personal 15% hurdle and the buy price drops to $234 low, $347 middle, $514 high.

So I raised my watch-list alert from $330 to $350, because the math got better. If it falls to $350 or below, I'll get the alert, rerun the numbers, and either buy the stock outright or sell cash-secured puts.

Up or Down, the Price Itself Tells You Nothing

Let me share the biggest lesson investing has taught me. The fact that a stock went up, or went down, tells you nothing on its own. What matters is what that price is buying you.

I held Intel all the way down to $17. It's at $130 today. And I don't believe Intel is worth $130. If you always think your stock is undervalued, you'll keep insisting it's cheap even after it's gone up 5x or 10x. Conclusions without a process eventually cost you your credibility.

So I say the same thing every time: don't copy my buys, or anyone's on the internet — not even Warren Buffett's. Copy the process of judging price against value instead. When you buy a stock, you're buying a piece of a business, and you're buying it for years from now, not for tomorrow.

FAQ

Q: Is Microsoft down 20% this year because of weak results? A: No. Cloud revenue grew 30% and margins are near record highs. The drop is driven by anxiety over when $190 billion of AI spending will pay off, and fear that AI agents will eat into software subscriptions.

Q: Isn't 22x earnings and 38x cash flow expensive? A: You can't judge on the multiple alone. The same multiple is cheap if the company grows double digits and expensive if it shrinks. Analysts expect Microsoft's revenue and earnings to more than double over seven years.

Q: Should I buy it now? A: That depends on your own required return. My 9% intrinsic value midpoint is $545, but at my personal 15% hurdle the buy price is around $347. I've set $350 as my alert and I'm waiting.